Should a researcher use AI tools for research?
It’s a confusing question. The instant answer should be: why not?
We all know that AI tools can help us in many ways, particularly in collecting information, conducting research, organising our thoughts and ideas, and shaping our writing. AI tools are particularly important for researchers whose native language is not English, as AI can help them improve the readability of a manuscript. It is now common for researchers to use several AI tools in their research; however, not all use AI in the same way. This situation leads to another question: When should a researcher use AI tools?
The primary concern for researchers, particularly on ethical grounds, is the boundary of AI use. How much help can we take from AI? If I instruct an AI tool to collect journal articles on my topic, that is acceptable. If I instruct an AI tool to proofread my English, is that permissible? But, if I instruct an AI tool to summarise an article, is that acceptable? Nowadays, if a person provides the main idea and a dataset to AI tools for writing a manuscript, the tools can easily write a manuscript in a short time. In that case, can AI claim the authorship? Can a non-human claim authorship?
Before discussing whether a researcher should use AI tools, I keep in mind that some people are eager to adopt new tools or technologies, while many have reservations. People sometimes even wonder whether a new technology will limit the scope of work. When the world started using computers, many thought people would lose their jobs due to this technological advancement. However, the scenario is that computer and internet technology actually open many new doors for improvement.
When I used AI for the first time, I had the same question: Is it going to replace some jobs regarding research, for instance, a data analyst or a proofreader? I instructed some AI tools to determine whether they correctly summarise a paper or to collect the relevant documents from the internet. After some research, I get the impression that AI can do some good jobs if the instructions are given correctly. AI needs to understand what a person wants, and if it can get a clear idea from the prompts, it can provide better output. This brings me another question to my mind — if we use AI, will it limit our thinking process? I consider this question to be the most important. Nothing should replace humans’ thinking process; if it did, humans would eventually halt the entire development process. This thought prompts the title question: Should a researcher use AI tools for research?
Clearly, there is no specific yes-or-no answer, because the term ‘use’ needs to be defined before getting a particular answer.
I was also checking what quality journals say about the use of AI tools in preparing a manuscript. The International Journal of STEM Education, published by Springer, indicates that it uses AI software to suggest improvements to the language quality of the manuscript. This statement demonstrates that using AI to improve a journal manuscript’s language is acceptable.
Another journal named Computers & Education, published by Elsevier, indicated that authors need to declare the use of AI as “Elsevier recognizes the potential of generative AI and AI-assisted technologies (“AI Tools”), when used responsibly, to help researchers work efficiently, gain critical insights fast and achieve better outcomes”. The journal also provided several guidelines to use AI and what to declare.
These two examples clearly indicate that using AI in the writing process is permitted; however, researchers must declare it in accordance with the journal’s policy. The important note is that writing a note is a part of the entire research process. How about using AI in other stages?
I read several articles online about the pros and cons of using AI tools in research. This reading creates the impression that researchers can use AI tools to support their work, particularly when the tools do not serve as a substitute for their own thinking. Researchers need to be very careful about all aspects and selective about specific parts.
When can a researcher use AI tools? Just some examples with caution
- Researchers can use AI tools to conduct an initial search to identify specific information relevant to their research. Even researchers can discuss complex ideas with AI to gain a clear understanding. The suggestion is not to rely on AI; a process of checking and rechecking is necessary. AI makes errors as humans do.
Get EdBrief’s stories in your inbox
Join Medium for free to get updates from this writer.
- If the idea generation or brainstorming involves human judgment or critical analysis, researchers can explore more ideas online with the help of AI. Still, it is the researchers’ responsibility to verify all information.
- Brainstorming and defining an idea are essential. Researchers can use AI to support the generation, refinement, selection, and justification of ideas. For this process, an AI tool can search online and collect some necessary data and information. Please be careful, as the AI-collected data may not be accurate. Researchers should recheck everything.
- AI tools can collect the specific literature from several journals and summarise it. To get the best results from AI, a researcher should provide a particular, clear prompt that specifies the information they are looking for, what to include, and what to exclude.
BIG NOTE
I have a personal reservation about summarising a paper by AI tools. Humans learn by using their brains, forming new neural connections in the process. Reading a paper, understanding its main points, critiquing the investigation, and getting a takeaway message from it should be the task of humans, not AI. Humans produce and receive knowledge, construct and deliver insights, and advance the world’s learning, development, and consciousness. If researchers hand over this process to AI, their brains will eventually become lazy, and their thinking will suffer. Research is about knowledge and insights; researchers can take help from AI but not appoint it to think on their behalf.
- Researchers can use AI tools for data cleaning and editing. While it is possible to obtain AI-generated analysis results, researchers should perform the analysis themselves. It’s also crucial on ethical grounds. Sharing research data with AI might clash with the ethical procedures.
- It is essential to express the ideas clearly. So, if researchers feel that AI tools can sharpen the statements, they can take this support. Again, researchers need to be very careful in taking any suggestions from AI tools. For instance, an AI tool can change the word from ‘output’ to ‘outcome’. Researchers need to decide which word best describes the statement. Sole reliance on AI tools can lead to disaster.
- Ensuring a language error-free document is important, particularly for publication purposes. Researchers can use AI tools for proofreading and minor language editing. Again, a conscious decision to take the AI suggestion is important.
- Researchers can use AI support for translation. Sometimes, it is essential to read a language that a researcher does not know. AI tools can help in this regard very quickly.
- Many researchers use software such as Microsoft Excel or SPSS to prepare graphs and figures. A prompt to AI tools can ease the researchers’ lives. AI can produce good visualisations, but researchers need to check whether AI tools provide the exact visualisations they want.
- There are so many administrative tasks involved in a research project. Researchers can use AI tools to organise them. They just should be careful that AI makes mistakes.
Three bottom lines
1. Researchers can take support from using AI tools, but always carefully.
2. Researchers must always be aware of whether AI is thinking on their behalf and making their decisions.
3. Researchers must grow a consciousness if they are becoming lazy in thinking and handing over the thinking process to AI tools.
